This approach has given Delphi Automotive the performance their users demanded while enabling incredible user access flexibility and enforcing strict data security perimeters. By placing the desktop close to the data instead of close to the engineer, we’ve created a fast and reliable user environment which can be presented to users around the globe. OST has engaged with Delphi Automotive to develop a high-performance desktop delivery environment, deploying workstation-class GPU-enabled Windows 10 desktops with high-bandwidth/low-latency connectivity to the telemetry “data lakes”. Delphi needed a more cost efficient and timely solution to access the data. Several hours per day were spent just in moving the data around the globe. This resulted in long lag times as data downloaded and uploaded. The data was stored in 1 of 2 data centers around the world, so the engineers had remote access to the information and were never directly connected to the content they needed to access. A ticket would be created for an incident that needed to be addressed, the engineer would download the data (which included full HD streaming video, graphics, and processor intensive workloads), make changes, and send the data and updated algorithms back to the data center. In order to analyze and make changes to the algorithms and program, engineers needed to be able to quickly access the data. One car would create about one terabyte of data a day.
When an autonomous car is deployed for testing, millions and millions of rows of data are collected on the performance of the system while in transit. This activity generates huge volumes of telemetry data from a wide range of sensors including high-definition cameras, LIDAR, RADAR, and more.
Delphi Automotive is pushing existing Active Safety technology forward in pursuit of an SAE Level 5 fully-autonomous self-driving vehicle.
0 kommentar(er)
